Search

Multiple kernel clustering (MKC) algorithm aims to group data into different categories by optimally integrating information from a group of pre-specified kernels. Though demonstrating superiorities in various applications, we observe that existing MKC algorithms usually do not sufficiently consider the local density around individual data samples and excessively limit the representation capacity of the learned optimal kernel, leading to unsatisfying performance. In this paper, we propose an algorithm, called optimal neighborhood MKC with adaptive local kernels (ON-ALK), to address the two issues. In specific, we construct adaptive local kernels to sufficiently consider the local density around individual data samples, where different numbers of neighbors are discriminatingly selected on each sample. Further, the proposed ON-ALK algorithm boosts the representation of the learned optimal kernel via relaxing it into the neighborhood area of weighted combination of the pre-specified kernels. To solve the resultant optimization problem, a three-step iterative algorithm is designed and theoretically proven to be convergent. After that, we also study the generalization bound of the proposed algorithm. Extensive experiments have been conducted to evaluate the clustering performance. As indicated, the algorithm significantly outperforms state-of-the-art methods in recent literatures on six challenging benchmark datasets, verifying its advantages and effectiveness. [ABSTRACT FROM AUTHOR]

Identifying the key risk factors of disease from a large amount of clinical data is a prerequisite for further scientific decision-making. In medical practice, the clinical symptom information of patients usually includes various types of data. Meanwhile, the occurrence and development of diseases are joint result of the mutual influence factors. Therefore, there is usually a correlation between attributes. In this paper, we discuss a kind of hybrid attribute feature selection problem considering the correlation between attributes. Firstly, we take the identification of disease pathogenic factors in medical decision as the background, and construct a hybrid attribute decision system. Secondly, by introducing kernel alignment, the uncertain relationship between attributes is defined. Based on this, a three-way clustering model in attribute space is established. Furthermore, a feature selection method for hybrid attribute data based on three-way clustering in attribute space is proposed. Finally, we applied the proposed model to identify the pathogenic factors of stroke and used 279 clinical random samples for simulation analysis. The results verified the applicability and validity of the model. The main contributions of this paper include two aspects. In terms of theory, by introducing kernel alignment, a three-way clustering algorithm in attribute space is established. Meanwhile, a hybrid attribute feature selection method based on three-way clustering is proposed. In terms of application, the proposed method is applied to identify risk factors of stroke. [ABSTRACT FROM AUTHOR]

• We propose a shape classification methodology using curvature-based features. • Our approach, termed EIFR, introduces an enhanced representation from a ReliefF-based feature selection. • EIFR also enhances the representation by introducing a kernel-alignment projection. • EIFR allows mapping relevant patterns on the input binary images and projecting them to a 2D space for visualization. • The enhanced feature representation favors both, shape recognition accuracy and data interpretability. [Display omitted] Shape classification from binary images is a challenging task within the computer vision community. Commonly, contour and structural features are computed to describe the objects and code patterns robust against rotation, scaling, and shape deformation. However, current techniques get a high-dimensional feature space decreasing the system performance and the attribute interpretability. Here, we introduce an enhanced and interpretable feature representation approach to support shape classification from binary images. Our method, named EIFR, employs a bag of contour fragments-based feature estimation, intrinsically robust to occlusion and shape deformation. Then, a ReliefF-based feature selection is applied to filter non-discriminative attributes. In turn, a kernel-alignment-based projection is used to measure the feature relevance enhancing the data representation through the matching between a similarity matrix computed from filtered attributes and a kernel matrix built from the shape labels. Attained results on benchmark datasets prove that EIFR improves the curvature-based features' interpretability and favors the classification performance. [ABSTRACT FROM AUTHOR]

The performance of most existing adaptive graph learning methods, which adjust data similarity matrix according to the data representation, depends on the hypotheses that the data representation is a good indicator of the underlying data structure. However, this hypothesis is not always applicable when dealing with high dimensional data. In this paper, we propose a novel kernel alignment unsupervised discriminative dimensionality reduction (KaUDDR) algorithm. By integrating adaptive graph learning and feature learning into a joint learning framework, graph construction and dimensionality reduction are conducted simultaneously to guarantee the optimality of graph for feature learning in the proposed algorithm. Data kernel and similarity indicator kernel are defined by learned graph and the projected data in a low-dimensional subspace, a compact and discriminative data representation in the projected subspace is obtained by means of kernel alignment to explore the consistency between the projected data kernel and similarity indicator kernel. Experimental results on dimensionality reduction as well as clustering show that our method consistently outperforms the related unsupervised dimensionality reduction algorithm. [ABSTRACT FROM AUTHOR]

In multi-label classification, multiple label variables in output space are equally important and can be predicted according to a common set of input variables. To improve the accuracy and efficiency of multi-label learner, measuring and utilizing label correlation is the core breakthrough. Extensive research on label correlation focuses on the co-occurrence or mutual exclusion frequency of label values in output space. In this paper, to handle the multi-label learning tasks, a novel method, named FL-MLC, is proposed by considering the influence of feature-label dependencies on inter-label correlations. In order to describe the intrinsic relationship between feature variable and label variable, the discriminant weight of any feature to label is first defined. Therefore, the concept of feature distribution for inputs on label is proposed to reflect the discriminant weights of features to the label. The corresponding calculation process is also designed based on multiple kernel learning and kernel alignment. Furthermore, the feature distributions on different labels are integrated into the feature distribution-based label correlation by using two different aggregation strategies. Obviously, arbitrary label variables with highly similar feature distributions have strong relevance. Thus, the feature distribution-based label correlation is applied to adjust the distance between the parameters for different labels in the predictive learner of FL-MLC method. Finally, the experimental results on twelve real-world datasets demonstrate that our methods achieves good effectiveness and versatility for multi-label classification. [ABSTRACT FROM AUTHOR]

Current studies have proposed the incorporation of kernel methods with graph representation learning, and graph kernels have attracted widespread attention for assessing graph similarity. However, owing to graph heterogeneity, developing appropriate kernels to extract complex structures and semantics from heterogeneous graphs is challenging. In this study, we developed a kernel-based heterogeneous graph neural network (GNN) model with a novel graph label kernel and an alignment-based aggregation mechanism. The graph label kernel captures the heterogeneous characteristics and integrates them with the graph aggregation mechanism into a unified framework. We employed the kernel alignment mechanism in a GNN framework to update node representation. The developed alignment-based GNN framework can automatically process feature propagation by modeling the similarity between node pairs and aligning the similarity matrix with the label kernel. Furthermore, model training avoids complex aggregation processes and captures distinguishable features during kernel alignment. Finally, we conducted a theoretical analysis to demonstrate the effectiveness of our method. Considering practicality, we optimized the implementation to reduce the running costs. We tested our method on three typical datasets, and the results showed that it outperformed state-of-the-art baselines. • Propose kernel-based GNN for heterogeneous graph learning. • Design novel graph kernel to capture heterogeneous features by kernel alignment. • Give an optimization implement to reduce running cost. • Provide theoretical proof and sufficient experiments to show the model validity. [ABSTRACT FROM AUTHOR]

Multiple kernel clustering (MKC) based on global structure alignment (GSA) has unified many existing MKC algorithms, and shown outstanding clustering performance. However, we observe that most of existing GSA-based MKC algorithms only maximally align global structure of data with an ideal similarity matrix, while ignoring the local geometrical structure hidden in data, which is regarded to be important in improving the clustering performance. To address this issue, we propose a global and local structure alignment framework for MKC (GLSAMKC) which well considers both the alignment between the global structure and local structure of data with the same ideal similarity matrix. To illustrate the effectiveness of the proposed framework, we instantiate two specific GLSAMKC-based algorithms by exploiting the local structure with local linear embedding and locality preserving projection, respectively. A two-step alternate iterative and convergent optimization algorithm is developed to implement the resultant optimization problem. Extensive experimental results on five benchmark data sets demonstrate the superiority of proposed algorithms compared with the many state-of-the-art MKC algorithms, indicating the effectiveness of the proposed framework.

• A three-level MKL framework for soil spectral analysis is proposed. • Multiple kernels maximize the information of each band and perform feature selection. • Multiple spectral and heterogeneous sources may also be appropriately combined. • The framework was applied to the LUCAS VNIR–SWIR topsoil database. • The developed models were more accurate and interpretable than the state-of-the-art. To ensure the sustainability of the soil ecosystem, which is the basis for food production, efficient large-scale baseline predictions and trend assessments of key soil properties are necessary. In that regard, visible, near-infrared, and shortwave infrared (VNIR–SWIR) spectroscopy can provide an alternative for the expensive wet chemistry. In this paper, we examined the application of the Multiple-Kernel Learning (MKL) approach to soil spectroscopy by integrating the information from heterogeneous features. In particular, the proposed three-level MKL framework acts in the following way: at the first level, it uses multiple kernels at each spectral feature (wavelength) to maximize the information of each band. At the second level, it performs implicit feature selection at the spectral source level, enabling it to provide interpretable results. Finally, at the third level of integration it combines the complementary information contained within a pool of spectral sources, each derived from its own set of pre-processing techniques. Additionally, at this stage, the proposed approach is also capable of fusing heterogeneous sources of information, such as auxiliary predictors, which can assist the spectral predictions. The experimental analysis was conducted using the pan-European LUCAS (Land Use/Cover Area frame statistical Survey) topsoil database, with a goal to predict from the VNIR–SWIR spectra the concentration of soil organic carbon (SOC), a key indicator for agricultural productivity and environmental resilience. The particle size distribution which describes the soil texture was selected as the set of auxiliary predictors. The proposed MKL framework was compared with other state-of-the-art approaches, and the results indicated that it attains the best performance in terms of accuracy, whilst at the same time producing interpretable results. [ABSTRACT FROM AUTHOR]

Multi-label learning deals with data sets in which each example is assigned with a set of labels, and the goal is to construct a learning model to predict the label set for unseen examples. Multi-label data sets share the same problems with single-label data sets that usually possess high-dimensional features and may exist redundancy features, which will influence the performance of the algorithm. Thus it is obviously necessary to address feature selection in multi-label learning. Meanwhile, information among labels play an important role in multi-label learning, thereby it is significance to measure information among labels in order to improve the performance of learning algorithms. In this paper, we introduce kernel alignment into multi-label learning to measure the consistency between feature space and label space by which features are ranked and selected. Firstly we define an ideal kernel in label space as a convex combination of ideal kernels defined by each label, and a linear combination of kernels where each kernel corresponds to a feature. Secondly, through maximizing the kernel alignment value between linear combination kernel and ideal kernel, both weights in the two defined kernels are learned in this process simultaneously, and the learned weights of labels can be employed as the degree of labeling importance regarded as a kind of information among labels. Finally, features are ranked according to their weights in linear combined kernel, and a proper feature subset consisting of top ranking features is selected. Thus a novel method of feature selection for multi-label learning is developed which can learn and address importance degree of labels automatically, and effectiveness of this method is demonstrated by experimental comparisons. [ABSTRACT FROM AUTHOR]

Multiple kernel clustering based on local kernel alignment has achieved outstanding clustering performance by applying local kernel alignment on each sample. However, we observe that most of existing works usually assume that each local kernel alignment has the equal contribution to clustering performance, while local kernel alignment on different sample actually has different contribution to clustering performance. Therefore this assumption could have a negative effective on clustering performance. To solve this issue, we design a multiple kernel clustering algorithm based on self-weighted local kernel alignment, which can learn a proper weight to clustering performance for each local kernel alignment. Specifically, we introduce a new optimization variable- weight-to denote the contribution of each local kernel alignment to clustering performance, and then, weight, kernel combination coefficients and cluster membership are alternately optimized under kernel alignment frame. In addition, we develop a three-step alternate iterative optimization algorithm to address the resultant optimization problem. Broad experiments on five benchmark data sets have been put into effect to evaluate the clustering performance of the proposed algorithm. The experimental results distinctly demonstrate that the proposed algorithm outperforms the typical multiple kernel clustering algorithms, which illustrates the effectiveness of the proposed algorithm. [ABSTRACT FROM AUTHOR]

Kernel based methods are increasingly being used for data modeling because of their conceptual simplicity and outstanding performance on many learning tasks. And kernel alignment, which is usually employed to select particular kernel for a learning task, is an effective quantity to measure the degree of agreement between a kernel and a learning task. However, the existing kernel alignment methods are usually developed for single-label classification problems. In this paper, we consider kernel alignment for multi-label learning to address the problem of kernel selection. Our basic idea is that, firstly an ideal kernel is presented in terms of multiple labels. Then kernel is selected by selecting the parameters of a linear combination of base kernels through maximizing the alignment value between the combined kernel and ideal kernel, and the selected kernel is employed in the binary relevance approach for multi-label learning to construct SVMs as classifiers. Furthermore, our proposed method is improved by considering local kernel alignment criterion. Our idea of selecting kernels by kernel alignment for multi-label learning is experimentally demonstrated to be effective in terms of classification accuracy. [ABSTRACT FROM AUTHOR]

Protein-DNA interactions are ubiquitous in a wide variety of biological processes. Correctly locating DNA-binding residues solely from protein sequences is an important but challenging task for protein function annotations and drug discovery, especially in the post-genomic era where large volumes of protein sequences have quickly accumulated. In this study, we report a new predictor, named TargetDNA, for targeting protein-DNA binding residues from primary sequences. TargetDNA uses a protein's evolutionary information and its predicted solvent accessibility as two base features and employs a centered linear kernel alignment algorithm to learn the weights for weightedly combining the two features. Based on the weightedly combined feature, multiple initial predictors with SVM as classifiers are trained by applying a random under-sampling technique to the original dataset, the purpose of which is to cope with the severe imbalance phenomenon that exists between the number of DNA-binding and non-binding residues. The final ensembled predictor is obtained by boosting the multiple initially trained predictors. Experimental simulation results demonstrate that the proposed TargetDNA achieves a high prediction performance and outperforms many existing sequence-based protein-DNA binding residue predictors. The TargetDNA web server and datasets are freely available at http://csbio.njust.edu.cn/bioinf/TargetDNA/ for academic use. [ABSTRACT FROM PUBLISHER]

Multiple Kernel Learning (MKL) literature has mostly focused on learning weights for base kernel combiners. Recent works using instance dependent weights have resulted in better performance compared to fixed weight MKL approaches. This may be attributed to the fact that, different base kernels have varying discriminative capabilities in distinct local regions of input space. We refer to the zones of classification expertize of base kernels as their “Regions of Success” (RoS). We propose to identify and model them (during training) through a set of instance dependent success prediction functions (SPF) having high values in RoS (and low, otherwise). During operation, the use of these SPFs as instance dependent weighing functions promotes locally discriminative base kernels while suppressing others. We have experimented with 21 benchmark datasets from various domains having large variations in terms of dataset size, interclass imbalances and number of features. Our proposal has achieved higher classification rates and balanced performance (for both positive and negative classes) compared to other instance dependent and fixed weight approaches. [ABSTRACT FROM AUTHOR]

To effectively learn heterogeneous features extracted from raw LiDAR point cloud data for landcover classification, a multiple kernel sparse representation classification (MKSRC) framework is proposed in this paper. In the MKSRC, multiple kernel learning (MKL) is embedded into sparse representation classification (SRC). The heterogeneous features are first extracted from the raw LiDAR point cloud data before classification. These features contain useful information from different dimensions, including single point features and neighbor features. Based on feature extraction, on the one hand, MKL is reasonably integrated into the SRC, namely, different base kernels that are constructed with each heterogeneous feature separately are utilized in the process of sparse representation. Furthermore, joint sparsity model is also introduced into the MKSRC framework and multiple kernel joint SRC (MKJSRC) is then proposed. On the other hand, improved kernel alignment (IKA) methods are proposed to more effectively determine the weights of base kernels in both of MKSRC and MKJSRC. Experiments are conducted on three real airborne LiDAR data sets. The experimental results demonstrate that MKSRC and MKJSRC frameworks can effectively learn the heterogeneous features for LiDAR point cloud classification and outperforms the other state-of-the-art sparse representation-based classifiers and the recent MKL algorithm. Moreover, the proposed IKA is helpful to better determine the “optimal” weights of the base kernels in both MKSRC and MKJSRC than in the existing kernel alignment method. [ABSTRACT FROM PUBLISHER]

Existing kernel-based correlation analysis methods mainly adopt a single kernel in each view. However, only a single kernel is usually insufficient to characterize nonlinear distribution information of a view. To solve the problem, we transform each original feature vector into a 2-dimensional feature matrix by means of kernel alignment, and then propose a novel kernel-aligned multi-view canonical correlation analysis (KAMCCA) method on the basis of the feature matrices. Our proposed method can simultaneously employ multiple kernels to better capture the nonlinear distribution information of each view, so that correlation features learned by KAMCCA can have well discriminating power in real-world image recognition. Extensive experiments are designed on five real-world image datasets, including NIR face images, thermal face images, visible face images, handwritten digit images, and object images. Promising experimental results on the datasets have manifested the effectiveness of our proposed method. [ABSTRACT FROM AUTHOR]

Detection of data structures in spectral clustering approaches becomes a difficult task when dealing with complex distributions. Moreover, there is a need of a real user prior knowledge about the influence of the free parameters when building the graph. Here, we introduce a graph pruning approach, termed Kernel Alignment based Graph Pruning (KAGP), within a spectral clustering framework that enhances both the local and global data consistencies for a given input similarity. The KAGP allows revealing hidden data structures by finding relevant pair-wise relationships among samples. So, KAGP estimates the loss of information during the pruning process in terms of a kernel alignment-based cost function. Besides, we encode the sample similarities using a compactly supported kernel function that allows obtaining a sparse data representation to support spectral clustering techniques. Attained results shows that KAGP enhances the clustering performance in most of the cases. In addition, KAGP avoids the need for a comprehensive user knowledge regarding the influence of its free parameters. [ABSTRACT FROM AUTHOR]

The success of kernel methods is very much dependent on the choice of kernel. Kernel design and learning a kernel from the data require evaluation measures to assess the quality of the kernel. In recent years, the notion of kernel alignment, which measures the degree of agreement between a kernel and a learning task, is widely used for kernel selection due to its effectiveness and low computational complexity. In this paper, we present an overview of the research progress of kernel alignment and its applications. We introduce the basic idea of kernel alignment and its theoretical properties, as well as the extensions and improvements for specific learning problems. The typical applications, including kernel parameter tuning, multiple kernel learning, spectral kernel learning and feature selection and extraction, are reviewed in the context of classification framework. The relationship between kernel alignment and other evaluation measures is also explored. Finally, concluding remarks and future directions are presented. [ABSTRACT FROM AUTHOR]

The identification of DNA-binding proteins (DBPs) has always been a hot issue in the field of sequence classification. However, considering that the experimental identification method is very resource-intensive, the construction of a computational prediction model is worthwhile. This study developed and evaluated a hybrid kernel alignment maximization-based multiple kernel model (HKAM-MKM) for predicting DBPs. First, we collected two datasets and performed feature extraction on the sequences to obtain six feature groups, and then constructed the corresponding kernels. To ensure the effective utilisation of the base kernel and avoid ignoring the difference between the sample and its neighbours, we proposed local kernel alignment to calculate the kernel between the sample and its neighbours, with each sample as the centre. We combined the global and local kernel alignments to develop a hybrid kernel alignment model, and balance the relationship between the two through parameters. By maximising the hybrid kernel alignment value, we obtained the weight of each kernel and then linearly combined the kernels in the form of weights. Finally, the fused kernel was input into a support vector machine for training and prediction. Finally, in the independent test sets PDB186 and PDB2272, we obtained the highest Matthew's correlation coefficient (MCC) (0.768 and 0.5962, respectively) and the highest accuracy (87.1% and 78.43%, respectively), which were superior to the other predictors. Therefore, HKAM-MKM is an efficient prediction tool for DBPs., (Copyright © 2022. Published by Elsevier Ltd.)

This paper presents a novel context-based scene recognition method that enables mobile robots to recognize previously observed topological places in known environments or categorize previously unseen places in new environments. We achieve this by introducing the Histogram of Oriented Uniform Patterns (HOUP), which provides strong discriminative power for place recognition, while offering a significant level of generalization for place categorization. HOUP descriptors are used for image representation within a subdivision framework, where the size and location of sub-regions are determined using an informative feature selection method based on kernel alignment. Further improvement is achieved by developing a similarity measure that accounts for perceptual aliasing to eliminate the effect of indistinctive but visually similar regions that are frequently present in outdoor and indoor scenes. An extensive set of experiments reveals the excellent performance of our method on challenging categorization and recognition tasks. Specifically, our proposed method outperforms the current state of the art on two place categorization datasets with 15 and 5 place categories, and two topological place recognition datasets, with 5 and 27 places. [ABSTRACT FROM PUBLISHER]

Abstract: Large amounts of histology images are captured and archived in pathology departments due to the ever expanding use of digital microscopy. The ability to manage and access these collections of digital images is regarded as a key component of next generation medical imaging systems. This paper addresses the problem of retrieving histopathology images from a large collection using an example image as query. The proposed approach automatically annotates the images in the collection, as well as the query images, with high-level semantic concepts. This semantic representation delivers an improved retrieval performance providing more meaningful results. We model the problem of automatic image annotation using kernel methods, resulting in a unified framework that includes: (1) multiple features for image representation, (2) a feature integration and selection mechanism (3) and an automatic semantic image annotation strategy. An extensive experimental evaluation demonstrated the effectiveness of the proposed framework to build meaningful image representations for learning and useful semantic annotations for image retrieval. [Copyright &y& Elsevier]

The increase in spatial and spectral resolution of the satellite sensors, along with the shortening of the time-revisiting periods, has provided high-quality data for remote sensing image classification. However, the high-dimensional feature space induced by using many heterogeneous information sources precludes the use of simple classifiers: thus, a proper feature selection is required for discarding irrelevant features and adapting the model to the specific problem. This paper proposes to classify the images and simultaneously to learn the relevant features in such high-dimensional scenarios. The proposed method is based on the automatic optimization of a linear combination of kernels dedicated to different meaningful sets of features. Such sets can be groups of bands, contextual or textural features, or bands acquired by different sensors. The combination of kernels is optimized through gradient descent on the support vector machine objective function. Even though the combination is linear, the ranked relevance takes into account the intrinsic nonlinearity of the data through kernels. Since a naive selection of the free parameters of the multiple-kernel method is computationally demanding, we propose an efficient model selection procedure based on the kernel alignment. The result is a weight (learned from the data) for each kernel where both relevant and meaningless image features automatically emerge after training the model. Experiments carried out in multi- and hyperspectral, contextual, and multisource remote sensing data classification confirm the capability of the method in ranking the relevant features and show the computational efficience of the proposed strategy. [ABSTRACT FROM AUTHOR]

Abstract: This paper focuses on developing a new framework of kernelizing Mahalanobis distance learners. The new KPCA trick framework offers several practical advantages over the classical kernel trick framework, e.g. no mathematical formulas and no reprogramming are required for a kernel implementation, a way to speed up an algorithm is provided with no extra work, the framework avoids troublesome problems such as singularity. Rigorous representer theorems in countably infinite dimensional spaces are given to validate our framework. Furthermore, unlike previous works which always apply brute force methods to select a kernel, we derive a kernel alignment formula based on quadratic programming which can efficiently construct an appropriate kernel for a given dataset. [Copyright &y& Elsevier]

Abstract: Existing multi-view learning focuses on the problem of how to learn from data represented by multiple independent sets of attributes (termed as multi-view data), and has been proved to bring an excellent performance. However, in general, we have only a single set of attributes (termed as single-view data) available. The goal of this paper is to employ the multi-view viewpoint to develop a multi-view kernel machine for such a single-view data. The key of doing so is to associate each learning machine with one kernel, take it as one view and thus form a set of learning machines from their corresponding kernels, as a result, a multi-view kernel machine can be developed by synthesizing them into a single learning framework. Further, in the two-view (two-kernel) case, we explore the relationship between the generalization ability of the proposed kernel machine and its associated kernels, in which with the kernel alignment (KA) as a correlation measure between kernels, it is found that superior performance of the proposed machine results from a weaker correlation between the constitutive kernels. To the best of our knowledge, both the multi-view learning on single-view data and the KA measure used here have not appeared in any literature. In practice, we take the kernel modified Ho–Kashyap with squared (KMHKS) approximation of the misclassification errors as a learning machine to develop a multi-view KMHKS (MultiV-KMHKS) on single-view data. [Copyright &y& Elsevier]

This paper describes the use of kernel methods to classify tissue samples using near-infrared spectra in order to discriminate between samples, either with or without elastane. The aim of this real-world study is to identify an alternative method to classify textile products using near-infrared (NIR) spectroscopy in order to improve quality control, and to aid in the detection of counterfeit garments. The principles behind support vector machines (SVMs), of which the main idea is to linearly separate data, are recalled progressively in order to demonstrate that the decision function obtained is a global optimal solution of a quadratic programming problem. Generally, this solution is found after embedding data in another space F with a higher dimension by the means of a specific non-linear function, the kernel. For a selected kernel, one of the most important and difficult subjects concerning SVM is the determination of tuning parameters. Generally, different combinations of these parameters are tested in order to obtain a machine with adequate classification ability. With the kernel alignment method used in this paper, the most appropriate kernel parameters are identified rapidly. Since in many cases, data are embedded in F, a linear principal component (PC) analysis (PCA) can be considered and studied. The main properties and the algorithm of k-PCA are described here. This paper compares the results obtained in prediction for a linear classifier built in the initial space with the PCs from a PCA and those obtained in F with non-linear PCs from a k-PCA. In the present study, even if there are potentially discriminating wavelengths seen on the NIR spectra, linear discriminant analysis and soft independent modelling of class analogy results show that these wavelengths are not sufficient to build a machine with correct generalisation ability. The use of a non-linear method, such as SVM and its corollary methods, kernel alignment and k-PCA, is then justified. [Copyright &y& Elsevier]